Skip to main content
ABC News
Democrats Don’t Need Trump’s Voters To Retake The House

Stop me if you’ve seen a headline (or five) that proclaims something along the lines of: “Most Trump voters still support Trump.” Typically, the article includes quotes from Trump voters in Pennsylvania or Michigan. Sometimes it revolves around polling showing people don’t “regret” voting for Trump. The takeaway is usually: Trump still has the support of his base, which means Democrats haven’t cracked the Trump nut yet.

But here’s the thing: Democrats don’t need to crack that nut by 2018; Trump can hang on to most — if not all — of his base, and Democrats could still clean up in the midterm elections. Checking in with Trump’s supporters is worthwhile. But don’t mistake their level of satisfaction for a political prediction.

Let’s start with the basic fact that Trump won just 45.9 percent of the vote in 2016. That doesn’t make his victory any less legitimate — winning (the Electoral College) with less than a majority is still winning — but Trump has a smaller base than every president elected since 1972, except for Bill Clinton in 1992. Trump voters are not a majority.

More importantly for the sake of 2018, they don’t represent the majority of voters in the majority of congressional districts. Trump won more than 50 percent in 205 of 435 districts. If House Republicans won every district where Trump won a majority in 2016 but lost every other one, Democrats would control 230 seats. Among seats won by a Republican in 2016, Trump fell short of a majority in 40 districts. Democrats need to win only 24 of those to win control of the House.

Of course, Democrats are unlikely to run the table in districts where Trump got less than a majority. He still won a plurality in 25 of those districts. And Democratic candidates probably won’t win every voter who cast a ballot for Hillary Clinton or a third-party candidate in 2016. The larger point is just that Republicans need more than Trump voters to hold onto the House.

Which is why judging the political climate by looking only at how Trump’s voters are feeling is misleading. Most presidents hold on to most of their base.

2010 Democrat 86% 14%
Republican 7 93
Third party*
2006 Democrat 93 7
Republican 16 84
Third party 74 26
2002 Democrat 88 12
Republican 14 86
Third party 76 24
1998 Democrat 82 18
Republican 9 91
Third party 36 57
1994 Democrat 84 16
Republican 11 89
Third party 32 68
1990 Democrat 81 19
Republican 37 63
Third party*
1986 Democrat 83 17
Republican 35 65
Third party*
1982 Democrat 89 11
Republican 31 69
Third party 76 24
The president’s party typically holds onto most of their voters in midterm elections

*This option was not included in the survey or lacked a sufficient number of respondents. Excludes voters who voted for third party candidates in the midterm elections.
The 2014 exit poll did not ask voters who they cast a vote for in the previous presidential election.

Source: The New York Times

But even a little erosion among base support can make a difference. Moreover, turnout matters, as does whether independent voters approve of the president’s job performance.

In the last three midterm wave elections (1994, 2006 and 2010) that resulted in the president’s party losing the House, for example, the president’s party won at least 84 percent of the president’s voters. But that wasn’t enough. In 1994, voters who cast a ballot for a third-party candidate in 1992 (mostly for Ross Perot) turned against the Democrats, going more than 2-to-1 for GOP House candidates. The 2004 election was close enough that Democrats holding on to a bit more of their voters in 2006 was enough to make huge gains. In 2010, poor turnout among then-President Barack Obama’s 2008 voters (though its effects are often overstated) hurt Democrats across the country.

In short, how independents vote in 2018 and who turns out will play roles just as big as that of how satisfied Trump voters are. Even if the latter are super happy with Trump, if everyone else is super unhappy, Democrats will likely do well.

The GOP’s problem again comes back to Trump’s base being relatively small to begin with compared to the base support of past presidents. The latest poll from YouGov, for example, shows 88 percent of respondents who said they voted for Trump approve of his job performance. But 88 percent of the 46 percent of 2016 voters who chose Trump is just 40 percent. Overall, the YouGov survey found 54 percent of registered voters disapprove of Trump so far. If every person who currently disapproves of Trump’s job performance voted against the Republican Party’s House candidates in 2018, Democrats would almost certainly take control of the chamber.1

Obviously, how someone feels about the president isn’t a perfect proxy for how they’ll vote in a House election. Fortunately for House Republicans, they’ll probably hold some voters who don’t like Trump. Candidate quality still matters, and there is still a small incumbent advantage in House elections. That’s how Republicans won the House popular vote by 1 percentage point in 2016 even as Trump was losing the popular vote.

That said, opinions of the incumbent president and House voting patterns have become more closely linked in recent midterms. The president’s party has lost at least 83 percent of voters who disapprove of the president’s job in every midterm since 1994. In none did the president’s party win more than 87 percent of those who approved of the president’s job.2 These statistics are not good news for Republicans if Trump’s current approval rating (40 percent among voters) and current disapproval rating (55 percent) holds through the midterm. Even if Trump’s Republican Party wins the recent high water mark of 87 percent of those who approve of the job the president is doing and loses only 83 percent of those who disapprove, Republicans would still lose the House popular vote by 7 percentage points.3 That could be enough for them to lose the House.

We still have well over a year until the midterm elections. President Trump’s approval rating may recover. But if Republicans want to hold onto the House in 2018, they had better hope it gets at least a little better — no matter how much Trump’s 2016 voters still approve of him.


  1. Assuming a vote pattern that followed Trump’s approval ratings exactly (39 percent approve and 54 disapprove) and that there was a uniform swing in every congressional district (i.e., every district shifted the same amount) of 13 percentage points from Trump’s 2 percentage point popular vote loss to a -15 percentage point net approval rating, Democrats would control 274 districts.

  2. In some, the president’s party won an even lower share of those who approve of the president’s job and lost an even larger share of those who approve of the president’s job.

  3. There was no sign in 2016 of the relationship between House and presidential vote patterns dissipating. The correlation between the House margin and presidential margin in congressional districts was +0.90.

Harry Enten was a senior political writer and analyst for FiveThirtyEight.